Executive Summary : | The investigator has proposed several methods for estimating power flow in micro-grids, including the 7 Point Estimation Method (7PEM), Guass Quadrature Based Probabilistic Load Flow (GQPLF) method, and efficient variants of Monte Carlo Simulation (MCS). However, these methods often overlook the importance of prior distributions in power system analysis. The author argues that the current situation in micro-grids, where loads and power generation are unpredictable random variables, requires probabilistic power flow (PRPF) analysis. PRPF calculates voltage, angle, and power flow probability distributions, and Monte Carlo Simulation (MCS) is the benchmark solution. However, these methods and MCS often overlook state variables as random variables. To address this issue, the author proposes Bayesian inference, which involves defining prior distributions over state variables and suggesting a probability distribution function that relates power supplied to state variables. This approach can provide a new source of information for MCS and APM simulation algorithms. The project aims to explore the application of Bayesian inference framework for PRPF analysis, develop approximate Bayesian likelihood-free inference approaches for micro-grid problems, and develop contingency analysis for microgrid problems. |